Wireless resource allocation method, device, base station equipment readable storage medium

Abstract

The application relates to a wireless resource allocation method and device, a base station device and a computer readable storage medium. The method comprises the following steps: acquiring terminal information and communication service information of each terminal in a target cell configured by a base station device; the target cell represents a cell for covering a ground area and a low-altitude area; inputting the terminal information and the communication service information into a pre-trained first allocation model to obtain BWP allocation information of the target cell, and allocating a first BWP to each terminal according to the BWP allocation information; determining a target terminal from each terminal according to communication quality information of each terminal after the first BWP is allocated; inputting the communication quality information and the communication service information of the target terminal into a pre-trained second allocation model to obtain BWP adjustment information of the target terminal, and allocating a second BWP to the target terminal according to the BWP adjustment information. The method can solve the problem of wireless resource allocation under double-carrier shared AAU deployment, thereby improving the accuracy of wireless resource allocation.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a wireless resource allocation method, apparatus, base station equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] With the development of communication technology, a low-altitude communication technology has emerged. This technology can use platforms such as drones and balloons flying at low altitudes for data transmission and communication. Low-altitude communication has advantages such as low latency, high flexibility and convenient deployment, and can solve communication problems in areas where existing communication networks are insufficient.

[0003] In traditional technologies, in order to maintain the stability of low-altitude communication, dual-carrier shared AAU deployment is usually used to deploy base stations, thereby achieving air and ground coverage of the communication network.

[0004] However, since drone users mainly focus on upstream services and ground users mainly focus on downstream services, the service needs of air and ground users differ greatly. Furthermore, the service needs of different low-altitude services and ground services also differ significantly, which often leads to problems in the allocation of wireless resources. Therefore, the current low-altitude communication technology deployed through dual-carrier shared AAU cannot accurately allocate wireless resources. Summary of the Invention

[0005] Therefore, it is necessary to provide a wireless resource allocation method, apparatus, base station equipment, computer-readable storage medium, and computer program product that can improve the accuracy of wireless resource allocation in response to the above-mentioned technical problems.

[0006] In a first aspect, this application provides a wireless resource allocation method applied to base station equipment, comprising:

[0007] The terminal information and communication service information of each terminal in the target cell configured by the base station equipment are obtained; the target cell refers to a cell used to cover the ground area and the low-altitude area.

[0008] The terminal information and the communication service information are input into a pre-trained first allocation model to obtain the BWP allocation information of the target cell, and a first BWP is allocated to each terminal according to the BWP allocation information;

[0009] Based on the communication quality information of each terminal after the first BWP is allocated, the target terminal is determined from each terminal;

[0010] The communication quality information and communication service information of the target terminal are input into the pre-trained second allocation model to obtain the BWP adjustment information of the target terminal, and a second BWP is allocated to the target terminal according to the BWP adjustment information.

[0011] In one of the embodiments, before the terminal information and the communication service information are input into the pre-trained first allocation model, the method further comprises: obtaining first sample terminal information and first sample communication service information of a first sample terminal in a sample cell, and obtaining target BWP allocation information of the sample cell; inputting the first sample terminal information and the first sample communication service information into the first allocation model to be trained to obtain first predicted BWP allocation information of the sample cell; and training the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information to obtain the pre-trained first allocation model.

[0012] In one of the embodiments, the target BWP allocation information of the sample cell is obtained by: obtaining candidate BWP allocation information of the sample cell; allocating a third BWP to each of the first sample terminals according to the candidate BWP allocation information to obtain comprehensive communication quality information corresponding to the candidate BWP allocation information; and obtaining the target BWP allocation information of the sample cell from the candidate BWP allocation information based on the comprehensive communication quality information corresponding to each of the candidate BWP allocation information.

[0013] In one of the embodiments, the pre-trained first allocation model is obtained by training the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information, which comprises: training the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information to obtain a first allocation model to be updated; obtaining second sample terminal information and second sample communication service information of a second sample terminal in the sample cell, inputting the second sample terminal information and the second sample communication service information into the first allocation model to be updated to obtain second predicted BWP allocation information of the sample cell; allocating a fourth BWP to each of the second sample terminals according to the second predicted BWP allocation information to obtain comprehensive communication quality information corresponding to the second predicted BWP allocation information; and updating the first allocation model to be updated by using the comprehensive communication quality information corresponding to the second predicted BWP allocation information to obtain the pre-trained first allocation model.

[0014] In one of the embodiments, before the communication quality information and the communication service information of the target terminal are input into the pre-trained second allocation model, the method further comprises: obtaining communication service information of a third sample terminal and first communication quality information of the third sample terminal; inputting the communication service information of the third sample terminal and the first communication quality information into the second allocation model to be trained, obtaining BWP adjustment action information of the third sample terminal through the second allocation model; obtaining second communication quality information of the third sample terminal after performing BWP adjustment action according to the BWP adjustment action information of the third sample terminal, obtaining reward information of the second allocation model based on the first communication quality information and the second communication quality information; training the second allocation model to be trained by using the reward information to obtain the pre-trained second allocation model.

[0015] In one of the embodiments, the training of the second allocation model to be trained by using the reward information to obtain the pre-trained second allocation model comprises: training the second allocation model to be trained by using the reward information to obtain a second allocation model to be evaluated; obtaining communication service information of a fourth sample terminal and third communication quality information of the fourth sample terminal; inputting the communication service information of the fourth sample terminal and the third communication quality information of the fourth sample terminal into the second allocation model to be evaluated to obtain BWP adjustment action information of the fourth sample terminal; obtaining fourth communication quality information of the fourth sample terminal after performing BWP adjustment action according to the BWP adjustment action information of the fourth sample terminal, and obtaining a model evaluation result of the second allocation model to be evaluated based on the fourth communication quality information; in a case where the model evaluation result meets a preset evaluation condition, taking the second allocation model to be evaluated as the pre-trained second allocation model.

[0016] In one of the embodiments, the terminal information comprises terminal types and terminal positions of the terminals, and the communication service information comprises service priorities and service demand bandwidths of the terminals; the obtaining of the terminal information and the communication service information of the terminals in the target cell comprises: obtaining terminal types, terminal positions, and communication service type information of the terminals; and obtaining service priorities and service demand bandwidths of the terminals according to the communication service type information.

[0017] In a second aspect, the application further provides a wireless resource allocation device applied to a base station device, comprising:

[0018] a terminal information obtaining module, configured to obtain terminal information and communication service information of terminals in a target cell configured by the base station device; the target cell represents a cell used for covering a ground area and a low-altitude area;

[0019] a first BWP allocation module configured to input the terminal information and the communication service information into a pre-trained first allocation model to obtain BWP allocation information of the target cell, and allocate a first BWP to each terminal according to the BWP allocation information;

[0020] a target terminal obtaining module configured to determine a target terminal from the terminals according to communication quality information of the terminals after the first BWP is allocated;

[0021] a second BWP allocation module configured to input the communication quality information of the target terminal and the communication service information into a pre-trained second allocation model to obtain BWP adjustment information of the target terminal, and allocate a second BWP to the target terminal according to the BWP adjustment information.

[0022] In a third aspect, the present application also provides a base station device, comprising a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method in any one of the embodiments of the first aspect when executing the computer program.

[0023] In a fourth aspect, the present application also provides a computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the method in any one of the embodiments of the first aspect.

[0024] In a fifth aspect, the present application also provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps of the method in any one of the embodiments of the first aspect.

[0025] The wireless resource allocation method, device, base station equipment, computer readable storage medium and computer program product described above, terminal information and communication service information of each terminal in a target cell configured by the base station equipment are obtained; the target cell represents a cell for covering a ground area and a low-altitude area; the terminal information and the communication service information are input into a pre-trained first allocation model to obtain BWP allocation information of the target cell, and a first BWP is allocated to each terminal according to the BWP allocation information; a target terminal is determined from each terminal according to communication quality information of each terminal after the first BWP is allocated; the communication quality information of the target terminal and the communication service information are input into a pre-trained second allocation model to obtain BWP adjustment information of the target terminal, and a second BWP is allocated to the target terminal according to the BWP adjustment information. The application obtains terminal information and communication service information of each terminal in a target cell covering a ground area and a low-altitude area through a base station equipment, so as to input the terminal information and the communication service information into a pre-trained first allocation model to obtain BWP allocation information of the target cell, so as to allocate a first BWP to each terminal of the target cell according to the BWP allocation information. Then, a target terminal which needs to adjust the BWP allocation again can be determined according to communication quality information of each terminal after the first BWP is allocated, so as to input the communication quality information of the target terminal and the communication service information into a pre-trained second allocation model to obtain BWP adjustment information of the target terminal, so as to allocate a second BWP to the target terminal. In this way, the allocation of the wireless resource BWP is completed from the whole target cell and an individual target terminal, and the target cell can cover a ground area and a low-altitude area, so as to solve the problem of wireless resource allocation under the deployment of double-carrier shared AAU, thereby improving the accuracy of wireless resource allocation. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related art, the drawings needed to be used in the description of the embodiments of the present application or the related art will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other related drawings can be obtained by those skilled in the art without any creative effort.

[0027] Figure 1 An application environment diagram of a wireless resource allocation method in an embodiment;

[0028] Figure 2 A flowchart of a wireless resource allocation method in an embodiment;

[0029] Figure 3 A flowchart of training a first allocation model in an embodiment;

[0030] Figure 4Flowchart of updating the first allocation model in one embodiment;

[0031] Figure 5 Flowchart of training the second allocation model in one embodiment;

[0032] Figure 6 Flowchart of evaluating the second allocation model in one embodiment;

[0033] Figure 7 Schematic diagram of carrier coverage of the space-to-ground base station and the ground-to-space base station in one embodiment;

[0034] Figure 8 Schematic diagram of selecting bandwidth resources according to service requirements in one embodiment;

[0035] Figure 9 Schematic diagram of BWP allocation of the space cell and the ground cell in one embodiment;

[0036] Figure 10 Flowchart of the space-to-ground integrated base station network dynamic resource allocation method in one embodiment;

[0037] Figure 11 Structural block diagram of the wireless resource allocation device in one embodiment;

[0038] Figure 12 Internal structure diagram of the base station device in one embodiment. DETAILED DESCRIPTION

[0039] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0040] The wireless resource allocation method provided by the embodiments of the present application can be applied to, for example, Figure 1The application environment shown. Among them, the base station device 101 communicates with the terminal device 102 in the low altitude area and the terminal device 103 in the ground area through wireless signals. Specifically, the base station device 101 can obtain the terminal information and communication service information of each terminal in the target cell configured by the base station device, that is, the terminal device 102 located in the low altitude area and the terminal device 103 located in the ground area, and then input the terminal information and communication service information into the pre-trained first allocation model to obtain the BWP allocation information of the target cell, so as to allocate the first BWP to each terminal, that is, to allocate the first BWP to each of the terminal device 102 and the terminal device 103. Then, according to the communication quality information of each terminal after allocating the first BWP, the target terminal can be screened out, and then the communication quality information and communication service information of the target terminal are input into the pre-trained second allocation model to obtain the BWP adjustment information for the target terminal, so as to realize the allocation of the second BWP to the target terminal. The base station device 101 can be an air-ground coverage base station, the terminal device 102 can be a communication device running in the cell in the low altitude area such as a drone, and the terminal device 103 can be but not limited to various personal computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices and other devices communicating in the ground area.

[0041] In one embodiment, as shown in Figure 2 , a wireless resource allocation method is provided, which is applied to the base station device 101 in Figure 1 for example, including the following steps:

[0042] Step S201, obtaining the terminal information and communication service information of each terminal in the target cell configured by the base station device 101; the target cell represents a cell for covering the ground area and the low altitude area.

[0043] Among them, the target cell refers to the cell configured by the base station device 101, which can include two parts, the ground cell for covering the ground area and the air cell for covering the low altitude area. In this embodiment, the base station device 101 can be an air-ground coverage base station, which can be deployed with two carrier signals, namely the first carrier and the second carrier, wherein the first carrier is deployed on the ground, and the second carrier realizes air-ground coverage, that is, part of the SSB beams of the second carrier cover the air, and part of the SSB beams cover the ground. The second carrier is configured as multiple different cells, which cover the low altitude and the ground respectively. The initial BWP of each cell is staggered in the frequency domain, and the dedicated BWP is dynamically allocated on demand.

[0044] The terminal information refers to terminal information of each terminal in the target cell configured by the base station device 101, such as the location of each terminal and the current BWP, and the communication service information refers to information used by each terminal to describe the communication service, such as the service type of the communication service.

[0045] Specifically, after the terminal accesses the target cell configured by the base station device, such as accessing the ground cell or the air cell in the air-ground coverage base station, the terminal information and the communication service information of the terminal can be sent to the base station device 101, and the base station device 101 can obtain the terminal information and the communication service information of each terminal in the target cell in real time.

[0046] In step S202, the terminal information and the communication service information are input into the pre-trained first allocation model to obtain BWP allocation information of the target cell, and each terminal is allocated a first BWP according to the BWP allocation information.

[0047] The first allocation model is a neural network model for overall BWP configuration of each terminal device in the target cell, and the BWP allocation information of the target cell is BWP allocation scheme information corresponding to the target cell. The allocation scheme can record the first BWP to be allocated to each terminal in the target cell. Specifically, the base station device 101 can input the obtained terminal information and communication service information into the pre-trained first allocation model, and output the BWP allocation information of the target cell from the first allocation model, so as to allocate a first BWP to each terminal in the target cell according to the BWP allocation information.

[0048] For example, the target cell includes terminal 1, terminal 2, and terminal 3, and the BWP includes BWP1, BWP2, and BWP3 of the ground cell, and BWP4, BWP5, and BWP6 of the air cell. The base station device 101 can input the terminal information and the communication service information into the pre-trained first allocation model to obtain the BWP allocation information of the target cell, which can be that terminal 1 is allocated to BWP1, terminal 2 is allocated to BWP3, and terminal 3 is allocated to BWP5. Then, each terminal in the target cell can be allocated a first BWP according to the above allocation information.

[0049] In step S203, a target terminal is determined from each terminal according to the communication quality information of each terminal after the first BWP is allocated.

[0050] The communication quality information can be used to represent the communication quality of each terminal in the target cell after the first BWP is allocated, and can be network throughput, user experience, energy efficiency, etc. The target terminal refers to a terminal device that needs to further adjust the BWP. After the first BWP allocation is completed by the first allocation model, since the BWP allocation information is based on the overall allocation information of the target cell, the target of the allocation is generally to reduce the optimal wireless network quality of the overall target cell. In this case, there may be a situation that the wireless network communication quality of part of the terminals is poor. Therefore, the base station device 101 can further determine the terminal device that needs to adjust the BWP as the target terminal based on the communication quality information fed back by each terminal after the first BWP is allocated.

[0051] In step S204, the communication quality information and the communication service information of the target terminal are input into the pre-trained second allocation model to obtain BWP adjustment information of the target terminal, and a second BWP is allocated to the target terminal according to the BWP adjustment information.

[0052] The second allocation model is a neural network model for BWP configuration of a single terminal device, which is mainly used to optimize the wireless resource configuration of a single terminal device. The BWP adjustment information of the target terminal refers to adjustment information for adjusting the BWP of the target terminal. Specifically, after the target terminal is determined, the base station device 101 can input the communication quality information and the communication service information of the target terminal into the pre-trained second allocation model, and output the BWP adjustment information of the target terminal from the second allocation model. Therefore, the base station device 101 can allocate a second BWP to the target terminal according to the BWP adjustment information.

[0053] In the above wireless resource allocation method, the terminal information and communication service information of each terminal in the target cell configured by the base station equipment 101 are obtained; the target cell refers to a cell used to cover the ground area and low-altitude area; the terminal information and communication service information are input into a pre-trained first allocation model to obtain the BWP allocation information of the target cell, and a first BWP is allocated to each terminal according to the BWP allocation information; based on the communication quality information of each terminal after the allocation of the first BWP, the target terminal is determined from each terminal; the communication quality information and communication service information of the target terminal are input into a pre-trained second allocation model to obtain the BWP adjustment information of the target terminal, and a second BWP is allocated to the target terminal according to the BWP adjustment information. This application obtains terminal information and communication service information of each terminal in a target cell that covers both ground and low-altitude areas through base station equipment 101. The terminal information and communication service information are then input into a pre-trained first allocation model to obtain BWP allocation information for the target cell. A first BWP is allocated to each terminal in the target cell according to the BWP allocation information. Afterwards, based on the communication quality information of each terminal after the first BWP allocation, the target terminals requiring further BWP allocation are determined. The communication quality information and communication service information of the target terminals are then input into a pre-trained second allocation model to obtain BWP adjustment information for the target terminals, thereby allocating a second BWP to the target terminals. This method achieves the allocation of radio resource BWPs from both the overall target cell and individual target terminals. Furthermore, the target cell can cover both ground and low-altitude areas, solving the problem of radio resource allocation under dual-carrier shared AAU deployment, thus improving the accuracy of radio resource allocation.

[0054] In one embodiment, such as Figure 3 As shown, before step S202, the following may also be included:

[0055] Step S301: Obtain the first sample terminal information and the first sample communication service information of the first sample terminal in the sample cell, and obtain the target BWP allocation information for the sample cell.

[0056] Here, sample cell refers to the cell sample used to train the first allocation model. Similar to the target cell, sample cell may also include ground-to-ground cells for covering the ground area and air-to-air cells for covering the low-altitude area. First sample terminal refers to the terminal used in the sample cell to train the first allocation model. This terminal can be a terminal within the coverage area of ​​the sample cell. First sample terminal information refers to the terminal information corresponding to the first sample terminal, and first sample communication service information refers to the communication service information corresponding to the first sample terminal.

[0057] The target BWP allocation information of the sample cell refers to the BWP allocation information that is optimal for the overall communication performance of the sample cell. Specifically, the terminal device 101 can also collect first sample terminal information of a first sample terminal in the sample cell and first sample communication service information, and obtain the BWP allocation information that is optimal for the overall communication performance of the sample cell as the target BWP allocation information of the sample cell.

[0058] In step S302, the first sample terminal information and the first sample communication service information are input into the first allocation model to be trained to obtain first predicted BWP allocation information of the sample cell.

[0059] The first predicted BWP allocation information is BWP allocation information for the first sample terminal in the sample cell output by the first allocation model. The first allocation model to be trained refers to a first allocation model that needs to be trained by a neural network model. Specifically, the base station device 101 can input the first sample terminal information and the first sample communication service information into the first allocation model to be trained, so as to output the first predicted BWP allocation information of the sample cell by the first allocation model to be trained.

[0060] In step S303, the first allocation model to be trained is trained according to the difference between the first predicted BWP allocation information and the target BWP allocation information to obtain a pre-trained first allocation model.

[0061] After obtaining the first predicted BWP allocation information, the difference between the first predicted BWP allocation information and the target BWP allocation information can be obtained. For example, a loss value between the first predicted BWP allocation information and the target BWP allocation information can be obtained by setting a loss function to represent the difference. Then, the first allocation model can be trained by using the above difference. For example, the pre-trained first allocation model can be obtained by continuously optimizing model parameters by using a back propagation algorithm to minimize the loss function.

[0062] In this embodiment, the training of the first allocation model can also be implemented by using a deep learning algorithm. The first sample terminal information of the first sample terminal in the sample cell, the first sample communication service information, and the target BWP allocation information of the sample cell are obtained, and the first allocation model is trained by using the above information. In this way, the accuracy of the predicted BWP allocation information for the overall cell output by the first allocation model can be improved.

[0063] Further, the step S201 can further include: obtaining candidate BWP allocation information of the sample cell; allocating the third BWP to each first sample terminal according to the candidate BWP allocation information, to obtain comprehensive communication quality information corresponding to the candidate BWP allocation information; and obtaining target BWP allocation information of the sample cell from the candidate BWP allocation information based on the comprehensive communication quality information corresponding to each candidate BWP allocation information.

[0064] The candidate BWP allocation information can refer to multiple selectable BWP allocation information of the sample cell, for example, multiple BWP allocation information generated by an optimization algorithm, and the comprehensive communication quality information corresponding to the candidate BWP allocation information can refer to overall communication quality information of the sample cell after the first sample terminal is allocated the third BWP according to the candidate BWP allocation information, for example, communication quality information of each first sample terminal after the first sample terminal is allocated the third BWP.

[0065] Specifically, before obtaining the target BWP allocation information of the sample cell, the base station device 101 can generate multiple candidate BWP allocation information of the sample cell by an optimization algorithm, and then allocate the third BWP to each first sample terminal according to each candidate BWP allocation information to obtain communication quality information of each first sample terminal under each candidate BWP allocation information, so as to determine comprehensive communication quality information corresponding to each candidate BWP allocation information. Then, the candidate BWP allocation information with the best comprehensive communication quality can be screened out as the target BWP allocation information of the sample cell according to the comprehensive communication quality information corresponding to each candidate BWP allocation information.

[0066] For example, the candidate BWP allocation information includes candidate BWP allocation information 1, candidate BWP allocation information 2, and candidate BWP allocation information 3. The base station device 101 can allocate the third BWP to each first sample terminal in the sample cell according to the above candidate BWP allocation information, and obtain comprehensive communication quality information corresponding to each candidate BWP allocation information, which is comprehensive communication quality information 1, comprehensive communication quality information 2, and comprehensive communication quality information 3, respectively. Then, the candidate BWP allocation information with the best comprehensive communication quality of the sample cell can be screened out according to the above comprehensive communication quality information. If the comprehensive communication quality information 2 represents the best comprehensive communication quality, the candidate BWP allocation information 2 can be used as the target BWP allocation information of the sample cell.

[0067] In this embodiment, a plurality of candidate BWP allocation information of the sample cell can also be generated, so as to allocate the third BWP to each first sample terminal in the sample cell by using each candidate BWP allocation information, to obtain the comprehensive communication quality information corresponding to each candidate BWP allocation information, and to screen the target BWP allocation information based on the comprehensive communication quality information corresponding to each candidate BWP allocation information. In this way, it can be ensured that the target BWP allocation information can be the optimal BWP allocation information of the sample cell, so as to ensure that the first allocation model can output the cell-comprehensive-optimal BWP allocation information.

[0068] In addition, as shown in Figure 4 Step S303 can further include:

[0069] Step S401, according to the difference between the first predicted BWP allocation information and the target BWP allocation information, training the first allocation model to be trained to obtain the first allocation model to be updated.

[0070] The first allocation model to be updated refers to the first allocation model that needs to be further optimized and updated after training. In this embodiment, after the base station device 101 completes the preliminary training of the first allocation model by using the difference between the first predicted BWP allocation information and the target BWP allocation information, the first allocation model to be updated can be obtained.

[0071] Step S402, obtaining the second sample terminal information and the second sample communication service information of the second sample terminal in the sample cell, and inputting the second sample terminal information and the second sample communication service information into the first allocation model to be updated to obtain the second predicted BWP allocation information of the sample cell.

[0072] The second sample terminal refers to the terminal device used for optimizing and updating the first allocation model in the sample cell, which can be the terminal device used in the actual test before the first allocation model is put into actual use. The second predicted BWP allocation information is the BWP allocation information output by the first allocation model to be updated for the second sample terminal in the sample cell.

[0073] Specifically, before the first allocation model is put into actual use, the first allocation model to be updated can also be subjected to actual network testing. At this time, the base station device 101 can collect the terminal information and the communication service information of the second sample terminal accessed during the test running as the second sample terminal information and the second sample communication service information. Then, the second sample terminal information and the second sample communication service information can be input into the first allocation model to be updated, so as to obtain the second predicted BWP allocation information of the sample cell.

[0074] Step S403, according to the second prediction BWP allocation information, a fourth BWP is allocated to each second sample terminal, and the integrated communication quality information corresponding to the second prediction BWP allocation information is obtained.

[0075] Step S404, the first allocation model to be updated is updated by using the integrated communication quality information corresponding to the second prediction BWP allocation information, and the pre-trained first allocation model is obtained.

[0076] After obtaining the second prediction BWP allocation information, the base station device 101 can allocate a fourth BWP to the second sample terminal by using the second prediction BWP allocation information, so as to obtain the communication quality information of each second sample terminal, and further determine the integrated communication quality information corresponding to the second prediction BWP allocation information. Then, the first allocation model can be further optimized and updated by using the integrated communication quality information, so as to obtain the pre-trained first allocation model for deployment in an actual network.

[0077] In the embodiment, after the preliminary training of the first allocation model is completed by using the difference between the first prediction BWP allocation information and the target BWP allocation information, the first allocation model can be further updated by using the second sample terminal information of the second sample terminal and the second sample communication service information. In this way, the accuracy of the prediction BWP allocation information for the entire cell output by the first allocation model can be further improved.

[0078] In one embodiment, as shown in FIG. 2, before step S204, the method can further include: Figure 5

[0079] Step S501, obtaining the communication service information of a third sample terminal and the first communication quality information of the third sample terminal.

[0080] The third sample terminal is a terminal used for training the second allocation model for BWP adjustment of a single terminal. The first communication quality information refers to the original communication quality information of the third sample terminal. After the third sample terminal accesses the base station device 101, the base station device 101 can further obtain the communication service information of the current third sample terminal and the communication quality information of the third sample terminal on the current BWP as the first communication quality information of the third sample terminal.

[0081] Step S502, inputting the communication service information of the third sample terminal and the first communication quality information into the second allocation model to be trained, and obtaining the BWP adjustment action information of the third sample terminal by using the second allocation model.

[0082] ​The second distribution model to be trained refers to a second distribution model that needs to be trained. In this embodiment, after obtaining the communication service information and the first communication quality information of the third sample terminal, the base station device 101 can input the communication service information and the first communication quality information into the second distribution model to be trained, and output the BWP adjustment action information for the third sample terminal from the second distribution model.

[0083] In step S503, the second communication quality information of the third sample terminal after performing the BWP adjustment action according to the BWP adjustment action information of the third sample terminal is obtained, and the reward information of the second distribution model is obtained based on the first communication quality information and the second communication quality information.

[0084] In step S504, the second distribution model to be trained is trained by using the reward information, and a pre-trained second distribution model is obtained.

[0085] The second communication quality information refers to the communication quality information of the third sample terminal after performing the BWP adjustment, and the reward information is the reward information output by the reward function based on the first communication quality information and the second communication quality information. For example, the reward information can be calculated according to the first communication quality information and the second communication quality information by defining the reward function, and then the second distribution model can be trained by using the reward information in the manner of reinforcement learning, so as to obtain the pre-trained second distribution model.

[0086] In this embodiment, the BWP adjustment action of the third sample terminal is output by the second distribution model, so as to adjust the BWP where the third sample terminal is located according to the adjustment action. Then, the reward information can be obtained according to the first communication quality information before adjustment and the second communication quality information after adjustment, so as to train the second distribution model by using the reward information. In this way, the second distribution model is trained in the manner of reinforcement learning, and the accuracy of the predicted BWP adjustment information for the terminal individual output by the second distribution model can be improved.

[0087] Further, as shown in FIG. 5, step S504 can further include: Figure 6

[0088] In step S601, the second distribution model to be trained is trained by using the reward information, and a second distribution model to be evaluated is obtained.

[0089] The second distribution model to be evaluated refers to a second distribution model that needs to be evaluated. In this embodiment, after completing the training of the second distribution model, the second distribution model needs to be tested and evaluated. Specifically, after the base station device 101 completes the training of the second distribution model by using the reward information, a second distribution model that needs to be evaluated can be obtained.

[0090] ​Step S602, obtaining the communication service information of the fourth sample terminal and the third communication quality information of the fourth sample terminal.

[0091] Step S603, inputting the communication service information of the fourth sample terminal and the third communication quality information of the fourth sample terminal into the second distribution model to be evaluated to obtain the BWP adjustment action of the fourth sample terminal.

[0092] The fourth sample terminal refers to a sample terminal used for evaluating the second distribution model. In this embodiment, the base station device 101 needs to obtain the communication service information of the fourth sample terminal and the current communication quality information of the fourth sample terminal as the third communication quality information of the fourth sample terminal when evaluating the second distribution model. Then, the communication service information and the third communication quality information of the fourth sample terminal can be input into the second distribution model to be evaluated, so as to obtain the BWP adjustment action of the fourth sample terminal.

[0093] Step S604, obtaining the fourth communication quality information of the fourth sample terminal after performing the BWP adjustment action according to the BWP adjustment action information of the fourth sample terminal, and obtaining the model evaluation result of the second distribution model to be evaluated based on the fourth communication quality information.

[0094] Step S605, in the case that the model evaluation result meets the preset evaluation condition, taking the second distribution model to be evaluated as the pre-trained second distribution model.

[0095] After the base station device 101 obtains the BWP adjustment action of the fourth sample terminal, the BWP of the fourth sample terminal can be adjusted according to the action, and the communication quality information of the fourth sample terminal after adjusting the BWP can be obtained as the fourth communication quality information. Then, the second distribution model can be evaluated by using the fourth communication quality information to obtain the model evaluation result of the second distribution model. If the model evaluation result meets the preset evaluation condition, i.e., the second distribution model passes the model evaluation, the base station device 101 can deploy the second distribution model to be evaluated as the pre-trained second distribution model.

[0096] In this embodiment, after the base station device 101 completes the training of the second distribution model, the second distribution model can be evaluated by using the fourth sample terminal to obtain the model evaluation result of the second distribution model. Only when the model evaluation result meets the preset evaluation condition, the second distribution model to be evaluated can be taken as the pre-trained second distribution model. The model evaluation can further ensure the BWP adjustment effect of the second distribution model on the independent terminal device.

[0097] In one embodiment, the terminal information includes terminal types of the terminals and terminal locations of the terminals, and the communication service information includes service priorities of the terminals and service demand bandwidths of the terminals; step S201 can further include: obtaining the terminal types, the terminal locations and the communication service type information of the terminals; and obtaining the service priorities and the service demand bandwidths of the terminals according to the communication service type information.

[0098] In this embodiment, the terminal information of each terminal can include two parts, a terminal type of the terminal, for example, whether the terminal is a low-altitude user terminal or a ground user terminal, and a location where the terminal is currently located, including a location and an altitude where the terminal is located, and the communication service information can include a time priority of a communication service, for example, an urgent task, an emergency task, a general task and a non-urgent task, and can further include a demand bandwidth of a service performed by the terminal, for example, an ultra-large bandwidth, a large bandwidth, a medium bandwidth and a small bandwidth.

[0099] Specifically, the base station device 101 can first obtain the terminal types, the terminal locations of the terminals in the target cell, and the service types of the communication services performed by the terminals, as the communication service type information of the terminals, and then can identify the service priorities and the service demand bandwidths of the terminals according to the communication service type information, which can be pre-constructed with a corresponding relationship between different service types and the service priorities and the service demand bandwidths, so as to identify the service priorities and the service demand bandwidths through the above corresponding relationship.

[0100] In this embodiment, the base station device 101 can also identify the terminal types, the terminal locations and the communication service types of the terminals in the target cell, so as to identify the service priorities and the service demand bandwidths of the terminals based on the communication service types, and in this way, the accuracy of the terminal BWP allocation can be improved.

[0101] In one embodiment, a method for dynamically allocating resources in an air-ground integrated base station network is also provided, which can be applied to an air-ground coverage base station. The first carrier in the base station is deployed for ground coverage, and the second carrier simultaneously implements coverage for the ground and the air, that is, part of the SSB beams of the second carrier cover the air and part of the SSB beams cover the ground. The second carrier can be configured as multiple different cells, which cover the low air and the ground respectively. The initial BWP of each cell is staggered in the frequency domain, and the dedicated BWP is dynamically allocated on demand. The air coverage cell and the ground coverage cell each contain part of the SSB beams. Figure 7As shown, the air-ground coverage base station forms Cell1 f1 by the first carrier to the ground, and forms Cell2 f2 and Cell3 f2 by the second carrier to the air at the same time. The ground coverage base station forms Cell1’ f1 and Cell2’ f2 by the first carrier and the second carrier to the ground. Specifically, Cell1 f1 and Cell1’ f1 and Cell2 f2, Cell2’ f2 frequency points mainly access the users of the large network, and the Cell3 f2 frequency point mainly accesses the users of the unmanned aerial vehicle. The specific resource allocation mode needs to consider the user service situation comprehensively.

[0102] On the basis of the above-mentioned networking scheme, the embodiment proposes a dynamic resource allocation method of an air-ground integrated base station network. By comprehensively considering different service demands, service demands of unmanned aerial vehicle users and ground users, and network resource conditions, joint optimization and allocation of time domain, frequency domain and space domain resources are realized.

[0103] The unmanned aerial vehicle user mainly uses the service of the above industry, and usually needs to perform video, point cloud and other large bandwidth data backhaul. At other times, in the case of no data transmission or only reporting state information, the bandwidth requirement is very small. Therefore, the demand for bandwidth is very different at different times. In addition, in the unmanned aerial vehicle emergency rescue service scenario, the network delay requirement is very high, and high-reliability communication guarantee needs to be provided. The ground user mainly uses the service of the below industry, and also exists the case that different services have different bandwidth delay requirements provided by the network. In addition, the number of ground users and unmanned aerial vehicle users may be distributed differently at different times, and the load conditions of the air carrier and the ground carrier may be overloaded at some times and underloaded at some times, causing resource waste. Reasonable access needs to be performed according to the flight height of the unmanned aerial vehicle and the location of the ground user. Therefore, reasonable joint resource allocation of time domain, frequency domain and space domain needs to be performed according to the service characteristics, so as to guarantee the communication performance of different users and improve the resource utilization rate.

[0104] In NR, the channel bandwidth spans a large range, and if the unmanned aerial vehicle UE performing different tasks always works on the full bandwidth, the base station resource is wasted. Different UEs can be configured with different BWP, and the bandwidth resource used by different unmanned aerial vehicle UEs is dynamically selected according to the service demand. For example, Figure 8As shown, for example: in the T0 period, the UE traffic load is large and not sensitive to delay (such as unmanned aerial power line inspection services, and agricultural and forestry plant protection related video backhaul services), and the system configures a large bandwidth BWP1 (BW is 40MHz, SCS is 15kHz) for the UE; in the T1 period, the traffic load is small (such as unmanned aerial sensor generates sensing results), and the system configures a small bandwidth BWP2 (BW is 10MHz, SCS is 15kHz) for the UE, which can achieve the purpose of reducing power consumption on the premise of meeting basic communication needs; in the T2 period, the UE may have burst delay-sensitive services (such as unmanned aerial emergency rescue related services), or find that the resources in the frequency band of BWP1 are scarce, and then switch to a new BWP3 (BW is 20MHz, SCS is 60kHz); in other time periods such as T3 and T4, the UE configures different BWP resources or switches between different BWPs according to real-time service needs.

[0105] Functionally, BWP is mainly divided into two categories, namely initial BWP and dedicated BWP. The initial BWP is mainly used for the UE to receive RMSI, OSI to initiate random access, etc. The dedicated BWP is mainly used for data service transmission. Figure 9 As shown, for the initial BWP, the embodiment can configure different frequencies for the air-to-space and air-to-ground cells of the second carrier. For the dedicated BWP, the unmanned aerial UE in the serving cell performs BWP operation, and up to 4 BWPs are configured by the higher layer. The embodiment proposes a deep reinforcement learning based air-ground integrated base station network dynamic resource allocation method according to actual service needs, user load and other comprehensive considerations, to realize the collaborative management of time domain, frequency domain and spatial domain resources, as shown in Figure 10 .

[0106] First, a large number of low-altitude unmanned aerial vehicle users and ground users apply for network access. In the case of sufficient network resources, low-altitude users are preferentially guaranteed to reside in low-altitude cells, and ground users access the first carrier and the second carrier of the ground coverage, while considering load balancing based on the number of users or PRB utilization. In the case of insufficient low-altitude and ground resources, in order to guarantee the user experience of low-altitude users and ground users, as well as the experience of high-priority services such as high reliability and low latency, a multi-dimensional collaborative resource allocation scheme is started.

[0107] 1. Classification strategy:

[0108] Low-altitude network deployment initially provides a fixed resource allocation selection strategy. Users are labeled as low-altitude users (user1), ground users (user2), etc.; service time priority is labeled as urgent (p1), emergency (p2), general (p3), and not urgent (p4); the bandwidth requirement of the service is divided into super large bandwidth (b1), large bandwidth (b2), medium bandwidth (b3), and small bandwidth (b4); and considering user location, height, and cell load information, the user is allocated time domain, frequency domain, and spatial domain resources. According to the type of service, the time priority, bandwidth selection, and beam resources are determined. The frequency domain resources of the user are configured with different BWP according to the corresponding bandwidth requirement, and BWP switching is performed when the service changes.

[0109] 2. Intelligent learning:

[0110] Through initial deployment, information such as throughput, rate, etc. corresponding to different resource allocation schemes can be collected, forming a large number of data sets. By establishing a deep learning prediction model, the globally optimal resource allocation scheme is obtained. First, based on Shannon's theorem C = Blog2(1 + S / N), considering the case of multiple users, multiple carriers, and multiple beams within the scheduling period, the network-level low-altitude user and ground user experience optimization is achieved by maximizing channel capacity.

[0111] 3. Dynamic update:

[0112] Based on the previous step, after allocating time-frequency-space resources to low-altitude users and ground users, there may be a case where the allocation scheme is globally optimal, but not optimal for a single user. In the case of no full scheduling of resources, further optimization can be performed using the remaining resources based on the above scheme, and the wireless resource scheme is further updated. Through reinforcement learning, a reward and punishment mechanism is set up to continuously explore the resource allocation scheme. A single-user objective function is set, such as user rate. The remaining time domain, frequency domain, and spatial domain resources are optimized, and the BWP resources are continuously updated and allocated according to the specific service priority. When the allocation scheme is updated, the performance is improved, and a reward is given; otherwise, a penalty is given. The model is continuously learned, reinforced, and updated to jointly optimize the network overall experience and single-user experience, and the optimal resource allocation scheme is obtained.

[0113] The specific implementation method can refer to the following example process:

[0114] (1) The process of using deep learning for mobile network time domain, frequency domain, and spatial domain resource allocation with global optimization as the target is as follows:

[0115] 1. Data preparation:

[0116] Collect historical data of the mobile network, including network topology, user demand, channel status, etc. Preprocess and feature engineer the data according to the demand to construct input features and labels.

[0117] 2. Model design:

[0118] Design a deep learning model, which can be a convolutional neural network (CNN), recurrent neural network (RNN), or attention mechanism (Attention) structure. Determine the structure of the input layer, hidden layer, and output layer, as well as the activation function, loss function, etc.

[0119] 3. Model training:

[0120] Input the prepared data into the deep learning model for training. Continuously optimize the model parameters through the backpropagation algorithm to minimize the loss function.

[0121] 4. Resource allocation decision:

[0122] Use the trained deep learning model to intelligently allocate time-domain, frequency-domain, and spatial-domain resources of the mobile network. According to the output results of the model, dynamically adjust the time slot, subcarrier, beam, and other resource allocation schemes.

[0123] 5. Model optimization:

[0124] Continuously collect data in the actual network and optimize and update the deep learning model according to the feedback information. Online learning algorithms such as incremental learning or transfer learning can be used to adapt to changes in the network environment.

[0125] 6. Performance evaluation:

[0126] Evaluate the performance of the optimized model, including network throughput, user experience, energy efficiency, and other indicators.

[0127] 7. Deployment and real-time tuning:

[0128] Deploy the optimized model to the actual mobile network and monitor the network status in real time to perform real-time tuning of resource allocation.

[0129] (2) Targeting single-user experience, use deep reinforcement learning to update the time-domain, frequency-domain, and spatial-domain resource allocation of the mobile network. The process can be divided into the following steps:

[0130] 1. Problem definition:

[0131] Determine the goal of resource allocation, such as maximizing user throughput, experience rate, minimizing interference between users, etc. Define the state space, action space, and reward function.

[0132] 2. State representation:

[0133] The state of the mobile network is represented as a feature vector. The state includes information such as network topology, channel state, user demand, base station status, etc. The state vector can be constructed using historical data.

[0134] 3. Action Selection:

[0135] The optimal action is selected from the action space using deep reinforcement learning algorithms such as Deep Q-Network (DQN) or policy gradient methods. The action can be time-domain resource allocation (e.g., adjusting transmission slots), frequency-domain resource allocation (e.g., bandwidth allocation), or spatial-domain resource allocation (e.g., beam selection), etc.

[0136] 4. Reward Calculation:

[0137] A reward function is defined to reflect the impact of each action on system performance. The reward function can consider network throughput, user experience, energy efficiency, etc.

[0138] 5. Environment Interaction:

[0139] In the simulated environment or real network, the selected action is executed and the feedback from the environment is observed. The parameters of the policy network are updated based on the feedback from the environment.

[0140] 6. Learning and Optimization:

[0141] The deep reinforcement learning algorithm is used to learn the optimal policy to maximize the cumulative reward. Classical reinforcement learning algorithms such as DQN, Deep Deterministic Policy Gradient (DDPG), etc. can be used.

[0142] 7. Model Evaluation:

[0143] The trained model is evaluated, including performance testing on the test data set and actual effect verification in the real network.

[0144] 8. Deployment and Optimization:

[0145] The trained model is deployed to the actual mobile network and online optimization and adjustment are made according to the changes in the real-time environment.

[0146] Through the embodiment, the air and the ground are simultaneously covered through the mobile cellular network, and the same frequency band can be used to realize that the two cells cover the ground and the low altitude respectively through the adaptive BWP method. Through the method, the same base station device can be used to realize the coverage of the air and the ground, save the device deployment cost, and through the scheduling of the BWP granularity, the resources can be fully utilized to improve the user experience. In addition, the embodiment can also classify the users according to the priority, allocate the time domain resources, the BWP and the beam resources to the users according to the service demand, and use the deep learning scheme to take the global network experience optimization as the target to perform the index prediction and the resource allocation, and use the deep reinforcement learning to take the single user experience optimization as the target to perform the real-time dynamic update of the resource allocation scheme. Through the method, the resource allocation can be dynamically adjusted according to the business scene requirement, and the balance between the global experience optimization and the single user experience optimization can be ensured.

[0147] It should be understood that, although each step in the flowchart involved in each embodiment as described above is shown in sequence according to the arrow, these steps are not necessarily executed in sequence according to the order of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0148] Based on the same inventive concept, the embodiments of the present application also provide a wireless resource allocation device for implementing the wireless resource allocation method described above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more wireless resource allocation device embodiments provided below can refer to the limitations of the wireless resource allocation method described above, which will not be described here.

[0149] In one embodiment, as shown in Figure 11 A wireless resource allocation device is provided, applied to a base station device, comprising: a terminal information acquisition module 1101, a first BWP allocation module 1102, a target terminal acquisition module 1103 and a second BWP allocation module 1104, wherein:

[0150] The terminal information acquisition module 1101 is configured to acquire terminal information and communication service information of each terminal in a target cell configured by the base station device; the target cell represents a cell for covering a ground area and a low altitude area;

[0151] The first BWP allocation module 1102 is configured to input the terminal information and the communication service information into a pre-trained first allocation model to obtain BWP allocation information of the target cell, and allocate a first BWP for each terminal according to the BWP allocation information.

[0152] The target terminal obtaining module 1103 is configured to determine a target terminal from the terminals according to the communication quality information of each terminal after the first BWP is allocated.

[0153] The second BWP allocation module 1104 is configured to input the communication quality information of the target terminal and the communication service information into a pre-trained second allocation model to obtain BWP adjustment information of the target terminal, and allocate a second BWP for the target terminal according to the BWP adjustment information.

[0154] In an embodiment, the wireless resource allocation apparatus further includes a first model training module configured to obtain first sample terminal information and first sample communication service information of a first sample terminal in a sample cell, and obtain target BWP allocation information for the sample cell; input the first sample terminal information and the first sample communication service information into a first allocation model to be trained to obtain first predicted BWP allocation information of the sample cell; and train the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information to obtain the pre-trained first allocation model.

[0155] In an embodiment, the first model training module is further configured to obtain candidate BWP allocation information of the sample cell; allocate a third BWP for each first sample terminal according to the candidate BWP allocation information to obtain comprehensive communication quality information corresponding to the candidate BWP allocation information; and obtain the target BWP allocation information for the sample cell from the candidate BWP allocation information based on the comprehensive communication quality information corresponding to each candidate BWP allocation information.

[0156] In an embodiment, the first model training module is further configured to train the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information to obtain a first allocation model to be updated; obtain second sample terminal information and second sample communication service information of a second sample terminal in the sample cell; input the second sample terminal information and the second sample communication service information into the first allocation model to be updated to obtain second predicted BWP allocation information of the sample cell; allocate a fourth BWP for each second sample terminal according to the second predicted BWP allocation information to obtain comprehensive communication quality information corresponding to the second predicted BWP allocation information; and update the first allocation model to be updated by using the comprehensive communication quality information corresponding to the second predicted BWP allocation information to obtain the pre-trained first allocation model.

[0157] In an embodiment, the wireless resource allocation apparatus further comprises a second model training module configured to: obtain communication service information of a third sample terminal and first communication quality information of the third sample terminal; input the communication service information of the third sample terminal and the first communication quality information of the third sample terminal into the second allocation model to be trained, and obtain BWP adjustment action information of the third sample terminal by using the second allocation model; obtain second communication quality information of the third sample terminal after performing the BWP adjustment action according to the BWP adjustment action information of the third sample terminal, and obtain reward information of the second allocation model based on the first communication quality information and the second communication quality information; and train the second allocation model to be trained by using the reward information to obtain the pre-trained second allocation model.

[0158] In an embodiment, the second model training module is further configured to: train the second allocation model to be trained by using the reward information to obtain the second allocation model to be evaluated; obtain communication service information of a fourth sample terminal and third communication quality information of the fourth sample terminal; input the communication service information of the fourth sample terminal and the third communication quality information of the fourth sample terminal into the second allocation model to be evaluated to obtain BWP adjustment action information of the fourth sample terminal; obtain fourth communication quality information of the fourth sample terminal after performing the BWP adjustment action according to the BWP adjustment action information of the fourth sample terminal, and obtain a model evaluation result of the second allocation model to be evaluated based on the fourth communication quality information; and in a case where the model evaluation result meets a preset evaluation condition, use the second allocation model to be evaluated as the pre-trained second allocation model.

[0159] In an embodiment, the terminal information comprises terminal types and terminal positions of the terminals, and the communication service information comprises service priorities and service demand bandwidths of the terminals; the terminal information obtaining module 1101 is further configured to obtain the terminal types, the terminal positions, and communication service type information of the terminals; and obtain the service priorities and the service demand bandwidths of the terminals according to the communication service type information.

[0160] The above modules of the wireless resource allocation apparatus can be realized by software, hardware, or a combination thereof. The above modules can be embedded in or independent of a processor in the base station device in a hardware form, or stored in a memory in the base station device in a software form, so as to be called and executed by the processor.

[0161] In an embodiment, a base station device is provided, which can be a terminal, and an internal structure diagram of the base station device can be as shown in FIG. 11. Figure 12The base station device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input apparatus. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the base station device is used to provide computing and control capabilities. The memory of the base station device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the base station device is used to exchange information between the processor and external devices. The communication interface of the base station device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be realized through WIFI, mobile cellular network, Near Field Communication (NFC) or other technologies. The computer program is executed by the processor to realize a wireless resource allocation method.

[0162] Those skilled in the art can understand that, Figure 12 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the base station device to which the scheme of the present application is applied. The specific base station device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0163] In one embodiment, a computer device is also provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to realize the steps in each of the above method embodiments.

[0164] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to realize the steps in each of the above method embodiments.

[0165] In one embodiment, a computer program product is provided, including a computer program, and the computer program is executed by a processor to realize the steps in each of the above method embodiments.

[0166] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0167] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiment methods. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. The non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. The volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, the RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., without being limited thereto.

[0168] The technical features of the above embodiments can be combined in any manner. To make the description concise, all possible combinations of the technical features in the above embodiments are not described, but as long as the combinations of the technical features do not exist contradictions, they should be considered as the scope of the present application.

[0169] The above-described embodiments are merely illustrative of several embodiments of the present application, which are described in more detail and in a specific manner, but should not be construed as limiting the scope of the patent of the present application. It should be noted that, for those of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A wireless resource allocation method, characterized by, The method is applied to a base station device, and comprises: obtaining terminal information and communication service information of each terminal in a target cell configured by the base station device; the target cell represents a cell for covering a ground area and a low-altitude area; inputting the terminal information and the communication service information into a pre-trained first allocation model to obtain BWP allocation information of the target cell, and allocating a first BWP to each terminal according to the BWP allocation information; determining a target terminal from the terminals according to communication quality information of the terminals after the first BWP is allocated; inputting the communication quality information and the communication service information of the target terminal into a pre-trained second allocation model to obtain BWP adjustment information of the target terminal, and allocating a second BWP to the target terminal according to the BWP adjustment information; before the terminal information and the communication service information are inputted into the pre-trained first allocation model, the method further comprises: obtaining first sample terminal information and first sample communication service information of a first sample terminal in a sample cell, and obtaining target BWP allocation information of the sample cell; inputting the first sample terminal information and the first sample communication service information into a first allocation model to be trained to obtain first predicted BWP allocation information of the sample cell; and training the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information to obtain the pre-trained first allocation model; before the communication quality information and the communication service information of the target terminal are inputted into the pre-trained second allocation model, the method further comprises: obtaining communication service information of a third sample terminal and first communication quality information of the third sample terminal; inputting the communication service information of the third sample terminal and the first communication quality information into a second allocation model to be trained to obtain BWP adjustment action information of the third sample terminal through the second allocation model; obtaining second communication quality information of the third sample terminal after a BWP adjustment action is performed according to the BWP adjustment action information of the third sample terminal, and obtaining reward information of the second allocation model based on the first communication quality information and the second communication quality information; and training the second allocation model to be trained by using the reward information to obtain the pre-trained second allocation model.

2. The method of claim 1, wherein, the target BWP allocation information of the sample cell is obtained by: obtaining candidate BWP allocation information of the sample cell; allocating a third BWP to each first sample terminal according to the candidate BWP allocation information to obtain comprehensive communication quality information corresponding to the candidate BWP allocation information; based on the comprehensive communication quality information corresponding to each candidate BWP allocation information, the target BWP allocation information of the sample cell is obtained from each candidate BWP allocation information.

3. The method of claim 1, wherein, the first allocation model to be trained is trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information to obtain the pre-trained first allocation model, comprising: training the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information, to obtain a first allocation model to be updated; obtaining second sample terminal information and second sample communication service information of a second sample terminal in the sample cell, and inputting the second sample terminal information and the second sample communication service information into the first allocation model to be updated, to obtain second predicted BWP allocation information of the sample cell; allocating fourth BWPs to each of the second sample terminals according to the second predicted BWP allocation information, to obtain comprehensive communication quality information corresponding to the second predicted BWP allocation information; updating the first allocation model to be updated by using the comprehensive communication quality information corresponding to the second predicted BWP allocation information, to obtain the pre-trained first allocation model.

4. The method of claim 1, wherein, The training the second allocation model to be trained by using the reward information, to obtain the pre-trained second allocation model, includes: training the second allocation model to be trained by using the reward information, to obtain a second allocation model to be evaluated; obtaining communication service information of a fourth sample terminal and third communication quality information of the fourth sample terminal; inputting the communication service information of the fourth sample terminal and the third communication quality information of the fourth sample terminal into the second allocation model to be evaluated, to obtain a BWP adjustment action of the fourth sample terminal; obtaining fourth communication quality information of the fourth sample terminal after performing the BWP adjustment action according to the BWP adjustment action information of the fourth sample terminal, and obtaining a model evaluation result of the second allocation model to be evaluated based on the fourth communication quality information; in a case where the model evaluation result meets a preset evaluation condition, taking the second allocation model to be evaluated as the pre-trained second allocation model.

5. The method according to any one of claims 1 to 4, characterized in that, The terminal information includes terminal types and terminal positions of each of the terminals, and the communication service information includes service priorities and service demand bandwidths of each of the terminals; The obtaining the terminal information and the communication service information of each terminal in the target cell includes: obtaining terminal types, terminal positions, and communication service type information of each of the terminals; obtaining service priorities and service demand bandwidths of each of the terminals according to the communication service type information.

6. A wireless resource allocation apparatus, characterized by comprising: The device is applied to a base station equipment, and the device includes: a terminal information obtaining module, configured to obtain terminal information and communication service information of each terminal in a target cell configured by the base station equipment; the target cell represents a cell used for covering a ground area and a low-altitude area; a first BWP allocation module, configured to input the terminal information and the communication service information into a pre-trained first allocation model, to obtain BWP allocation information of the target cell, and allocate first BWPs to each of the terminals according to the BWP allocation information; a target terminal obtaining module, configured to determine a target terminal from each of the terminals according to communication quality information of each of the terminals after the first BWPs are allocated. The second BWP allocation module is configured to input the communication quality information and the communication service information of the target terminal into a pre-trained second allocation model to obtain BWP adjustment information of the target terminal, and allocate a second BWP to the target terminal according to the BWP adjustment information. The first model training module is configured to obtain first sample terminal information and first sample communication service information of a first sample terminal in a sample cell, and obtain target BWP allocation information of the sample cell; input the first sample terminal information and the first sample communication service information into a first allocation model to be trained to obtain first predicted BWP allocation information of the sample cell; and train the first allocation model to be trained according to a difference between the first predicted BWP allocation information and the target BWP allocation information to obtain the pre-trained first allocation model. The second model training module is configured to obtain communication service information of a third sample terminal and first communication quality information of the third sample terminal; input the communication service information of the third sample terminal and the first communication quality information into a second allocation model to be trained to obtain BWP adjustment action information of the third sample terminal through the second allocation model; obtain second communication quality information of the third sample terminal after performing a BWP adjustment action according to the BWP adjustment action information of the third sample terminal, and obtain reward information of the second allocation model based on the first communication quality information and the second communication quality information; and train the second allocation model to be trained by using the reward information to obtain the pre-trained second allocation model.

7. The apparatus of claim 6, wherein, The first model training module is further configured to obtain candidate BWP allocation information of the sample cell; allocate a third BWP to each of the first sample terminals according to the candidate BWP allocation information to obtain comprehensive communication quality information corresponding to the candidate BWP allocation information. The target BWP allocation information of the sample cell is obtained from the candidate BWP allocation information based on the comprehensive communication quality information corresponding to each of the candidate BWP allocation information. 8.A base station device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 5.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 5.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 5.

Images